Humanitarian Response

TYPHOON DAMAGE PREDICTION IN PHILIPPINES

“Towards a Global Impact-based Forecasting Model for Tropical Cyclones”. Natural Hazards and Earth System Sciences Journal, 2023

Forooshani M K, van den Homberg M, Kalimeri K, Kaltenbrunner A, Mejova Y, Milano L, Ndirangu P, Paolotti D, Teklesadik A, and Turner M

We improve an existing anticipatory action model for the Philippines by transforming the target variable into a high-resolution grid of 0.1 degrees, using only globally-available features, and introducing a two-stage model to focus on high-damage areas. We demonstrate that employing only the global features does not significantly influence model performance. Despite excluding local data on physical vulnerability and storm surge susceptibility, the two-stage model improves upon the municipality-based model with local features. When applied to anticipatory action, our two-stage model would show a higher True Positive rate, a lower False Negative rate, and an improved False Positive rate, implying that fewer resources would be wasted in anticipatory action. 

TYPHOON DAMAGE PREDICTION IN PHILIPPINES

“Towards a Global Impact-based Forecasting Model for Tropical Cyclones”. Natural Hazards and Earth System Sciences Journal, 2023

Forooshani M K, van den Homberg M, Kalimeri K, Kaltenbrunner A, Mejova Y, Milano L, Ndirangu P, Paolotti D, Teklesadik A, and Turner M

We improve an existing anticipatory action model for the Philippines by transforming the target variable into a high-resolution grid of 0.1 degrees, using only globally-available features, and introducing a two-stage model to focus on high-damage areas. We demonstrate that employing only the global features does not significantly influence model performance. Despite excluding local data on physical vulnerability and storm surge susceptibility, the two-stage model improves upon the municipality-based model with local features. When applied to anticipatory action, our two-stage model would show a higher True Positive rate, a lower False Negative rate, and an improved False Positive rate, implying that fewer resources would be wasted in anticipatory action. 

FORECASTING FOOD INSECURITY

“On the forecastability of food insecurity”. 2022

Foini P, Tizzoni M, Paolotti D, Omodei E

In this study, we tackle the problem of forecasting the sub-national daily evolution of the prevalence of people with insufficient food consumption. This metric, characterizing a given area at at given time, is obtained as the prevalence of households, in the specified area and time, that are measured to have poor or borderline food consumption according to one of the core household food security indicators, namely the Food Consumption Score (FCS), which captures households' dietary diversity and nutrient intake. Our study comes with a simple but fundamental message for governments and humanitarian organizations on the power of the data they collect: collecting data on a regular basis for long enough periods of time and across enough different geographic areas can not only allow to monitor the evolution of the situation in near real-time but also to inform forecasting models that would allow to produce estimates of how the situation is likely to evolve in the near future.

INFORMATION EXTRACTION FROM HUMANITARIAN REPORTS

“Developing natural language processing resources for the humanitarian sector” 2022

Tamagnone N, Kalimeri K, Mejova Y, Schifanella R, Cattuto C

The DEEP Platform aims to centralize, accelerate, and strengthen inter-agency response to humanitarian crises. We developed NLP tools for processing of humanitarian reports and extracting relevant information for analysis, visualization, and summary reporting from a wide range of sources .pdf/.doc/.docx/webpage/.html 

Open source code: https://github.com/the-deep/deepex

STRENGTHS AND LIMITATIONS OF RELATIVE WEALTH INDICES DERIVED FROM BIG DATA

Sartirano D, Kalimeri K, Cattuto C, Delamònica E, Garcia-Herranz M, Mockler A, Paolotti D, Schifanella R

Accurate relative wealth estimates in Low and Middle Income Countries (LMICS) are crucial to help policymakers address socio-demographic inequalities under the guidance of the Sustainable Development Goals set by the United Nations. 

In this work, we focus on the case of Indonesia, and examine one frontier-data derived Relative Wealth Index (RWI), created by the Facebook Data for Good initiative, that utilizes connectivity data from the Facebook Platform and satellite imagery data to produce a high-resolution estimate of wealth for 135 countries. We examine it in relation to asset-based relative wealth indices estimated from existing high quality national level traditional survey instruments, the USAID-developed Demographic Health Survey (DHS) and the Indonesian National Socio-economic survey (SUSENAS).